Sensor Fusion Integrity Evaluation for Autonomous Driving

The patent applies preliminary action by determining evaluation values for detection conditions before performing sensor data fusion. The system evaluates detection conditions (such as weather, lighting, sensor status) in advance and stores these evaluation values, which are then used during the fusion process to weight sensor data appropriately. This prevents sensor deficiencies from compromising the surroundings model while maintaining processing efficiency.

The patent applies local quality by assigning different evaluation values to different detection conditions and sensor modalities based on their specific characteristics and current state. Each sensor's data is weighted according to its own detection condition evaluation (e.g., camera data weighted differently in fog vs. clear weather, radar weighted differently in various scenarios). This allows the system to handle multiple sensor modalities effectively by treating each sensor's contribution locally and specifically rather than applying a uniform fusion approach.

The patent applies self-service by having the sensor system automatically evaluate its own detection conditions and generate evaluation values without requiring external intervention. The system autonomously monitors its own operational state, sensor deficiencies, and environmental conditions, then uses this self-generated information to weight and fuse sensor data. This reduces operational complexity while significantly improving measurement accuracy by preventing false positives and negatives.